Enhance shape to domain Conversion to Retain All Subpolygons

[Jinnkunn]

Currently, when converting a shapefile containing multiple subpolygons into a domain, only one subpolygon is retained, instead of the full set of polygons. It would be beneficial to further encapsulate shape and domain to support complete multi-polygon information.

[Jinnkunn]

This issue can be reproduced when exporting data for Alex's data request. The problem arises because an additional step is required: creating a convex_hull before instantiating a domain. For example:

gdf = gpd.read_file('./shape/')
gdf = gdf.explode(index_parts=True)  # index_parts=True to handle the index correctly
print(gdf.crs)

minx, miny, maxx, maxy =  gdf.iloc[0]['geometry'].bounds

domain_list = []
for idx in range(gdf.shape[0]):
    polygon_utm = gdf.iloc[idx]['geometry']
    domain_list.append(polygon_utm)
    minx, miny, maxx, maxy = min(minx, polygon_utm.bounds[0]), min(miny, polygon_utm.bounds[1]), max(maxx, polygon_utm.bounds[2]), max(maxy, polygon_utm.bounds[3])

new_geo_series = gpd.GeoSeries(shapely.ops.unary_union(domain_list))
new_geo_series = new_geo_series.convex_hull
mask = new_geo_series

domain = aisdb.gis.Domain(name=f'domain', zones=[{
        'name': f'domain',
        'geometry': mask.geometry.iloc[0]
    }  ])

Users unfamiliar with the Shapely library may find this process confusing. To improve usability, we could implement a function where users simply pass in the Shapely instance, and the function handles the creation of required geometry instances. Additionally, it could automatically convert the Coordinate Reference System (CRS) if the original CRS is not EPSG:4326.

This would eliminate the need for users to manually manipulate Shapely instances, making the workflow more intuitive and efficient.